CMPU4091 Visualizing Data Assignment

Tech companies with good benefits improve mental health of employees

Introduction

Mental health has become an increasingly important topic in modern workplaces, especially in high-stress and fast-paced industries like technology. Employees in tech companies often face long hours, tight deadlines, and a culture that prizes productivity—conditions that can exacerbate mental health challenges. This project explores the relationship between workplace benefits and mental health outcomes for employees in the tech sector using real-world survey data.

The primary goal of this analysis is to determine whether supportive workplace policies such as remote work, access to mental health care options, wellness programs, and flexible leave policies correlate with improved mental health outcomes. Specifically, we examine whether these benefits increase the likelihood of employees seeking treatment and reduce self-reported work interference due to mental health conditions.

The dataset used for is the analysis a 2014 Mental Health in Tech Survey found on kaggle (https://www.kaggle.com/datasets/osmi/mental-health-in-tech-survey). Transformations on gender and age were done, as well as an age group was added for better analysis.

• Some of the following columns are:

  • family_history: Do you have a family history of mental illness?

  • treatment: Have you sought treatment for a mental health condition?

  • work_interfere: If you have a mental health condition, do you feel that it interferes with your work?

  • no_employees: How many employees does your company or organization have?

  • remote_work: Do you work remotely (outside of an office) at least 50% of the time?

  • tech_company: Does your employer provide mental health benefits?

  • care_options: Do you know the options for mental health care your employer provides?

  • wellness_program: Does your employer provide resources to learn more about mental health issues and how to seek help?

  • anonymity: Is your anonymity protected if you choose to take advantage of mental health or substance abuse treatment resources?

  • leave: How easy is it for you to take medical leave for a mental health condition?

  • mental_health_consequence: Do you think that discussing a mental health issue with your employer would have negative consequences?

  • phys_health_consequence: Do you think that discussing a physical health issue with your employer would have negative consequences?

  • coworkers: Would you be willing to discuss a mental health issue with your coworkers?

  • supervisor: Would you be willing to discuss a mental health issue with your direct supervisor(s)?

  • mental_health_interview: Would you bring up a mental health issue with a potential employer in an interview?

  • phys_health_interview: Would you bring up a physical health issue with a potential employer in an interview?

  • mental_vs_physical: Do you feel that your employer takes mental health as seriously as physical health?

  • obs_consequence: Have you heard of or observed negative consequences for coworkers with mental health conditions in your workplace?

Data Preparation and Cleaning

# Load and clean dataset
mental_health <- read_csv("survey.csv") %>%
  janitor::clean_names()

Rows: 1259 Columns: 27
── Column specification ────────────────────────────────────────────────────────
Delimiter: ","
chr  (25): Gender, Country, state, self_employed, family_history, treatment,...
dbl   (1): Age
dttm  (1): Timestamp

ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

cat(strwrap(toString(unique(mental_health$gender)), width = 80), sep = "\n")

Female, M, Male, male, female, m, Male-ish, maile, Trans-female, Cis Female, F,
something kinda male?, Cis Male, Woman, f, Mal, Male (CIS), queer/she/they,
non-binary, Femake, woman, Make, Nah, All, Enby, fluid, Genderqueer, Androgyne,
Agender, cis-female/femme, Guy (-ish) ^_^, male leaning androgynous, Man, Trans
woman, msle, Neuter, Female (trans), queer, Female (cis), Mail, cis male, A
little about you, Malr, p, femail, Cis Man, ostensibly male, unsure what that
really means

# Data cleaning and transformation
mental_health <- mental_health %>%
  filter(!is.na(age) & age > 15 & age < 100) %>%
  mutate(
    gender = str_to_lower(gender),
    gender = case_when(
      str_detect(gender, "female|f") ~ "Female",
      str_detect(gender, "male|m") ~ "Male",
      TRUE ~ "Other"
    ),
    treatment = ifelse(treatment == "Yes", "Treated", "Not Treated"),
    age_group = cut(
      age,
      breaks = c(seq(10, 60, by = 5), Inf),  # 10, 15, ..., 60, Inf
      labels = c("10–15", "16–20", "21–25", "26–30", "31–35", "36–40", "41–45", "46–50", "51–55", "56–60", "60+"),
      right = TRUE,
      include.lowest = TRUE
    )
  )

Gender has various different unique entries ranging from different ways of representing the two main sexes (M, m, Male, male) to Queer identities. There were some mispellings such as msle and mail. Some of the queer identities include Enby, Genderqueer, Agender, Androgyne, Trans-female. There are different spellings on the same identity i.e trans-female, Trans woman, and Female (trans). Some people also wrote Cis male or female cis. To be able to group programatically the different queer identities would be difficult and not scalable thus they will be grouped into other. Any correct spelling of male or female along with F or M will be grouped as male and female. This kind of problem could be avoided if the instead of taking a string input in the survey you used bulletpoints and if necessary an other text input.

I filtered for rows in the age range of 10 to 100. There were outliers such as 999999, -1, 5 and 8 which are invalid. The age_group column was added to explore how different age groups feel. The age groups are grouped in increments of 5 starting from 10 to 60+.

There is a states column which is states for the USA. However there are non US countires where there is no states value for them.

Data Overview

datatable(mental_health)

Data Exploration

Below is the unique values for each column and the count.

for (col in names(mental_health)) {
  if (col %in% c("timestamp", "comments")) {
    next
  }
  cat("Value counts for:", col, "\n")
  print(table(mental_health[[col]], useNA = "ifany"))
  cat("\n---------------------------\n\n")
}

Value counts for: age 

18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 
 7  9  6 16 21 51 46 61 75 71 68 85 63 67 82 70 65 55 37 43 39 33 33 21 20 28 
44 45 46 47 48 49 50 51 53 54 55 56 57 58 60 61 62 65 72 
11 12 12  2  6  4  6  5  1  3  3  4  3  1  2  1  1  1  1 

---------------------------

Value counts for: gender 

Female   Male  Other 
   247    994     10 

---------------------------

Value counts for: country 

             Australia                Austria                Belgium 
                    21                      3                      6 
Bosnia and Herzegovina                 Brazil               Bulgaria 
                     1                      6                      4 
                Canada                  China               Colombia 
                    72                      1                      2 
            Costa Rica                Croatia         Czech Republic 
                     1                      2                      1 
               Denmark                Finland                 France 
                     2                      3                     13 
               Georgia                Germany                 Greece 
                     1                     45                      2 
               Hungary                  India                Ireland 
                     1                     10                     27 
                Israel                  Italy                  Japan 
                     5                      7                      1 
                Latvia                 Mexico                Moldova 
                     1                      3                      1 
           Netherlands            New Zealand                Nigeria 
                    27                      8                      1 
                Norway            Philippines                 Poland 
                     1                      1                      7 
              Portugal                Romania                 Russia 
                     2                      1                      3 
             Singapore               Slovenia           South Africa 
                     4                      1                      6 
                 Spain                 Sweden            Switzerland 
                     1                      7                      7 
              Thailand         United Kingdom          United States 
                     1                    184                    746 
               Uruguay 
                     1 

---------------------------

Value counts for: state 

  AL   AZ   CA   CO   CT   DC   FL   GA   IA   ID   IL   IN   KS   KY   LA   MA 
   7    7  138    9    4    4   15   12    4    1   28   27    3    5    1   20 
  MD   ME   MI   MN   MO   MS   NC   NE   NH   NJ   NM   NV   NY   OH   OK   OR 
   8    1   22   20   12    1   14    2    3    6    2    3   57   27    6   29 
  PA   RI   SC   SD   TN   TX   UT   VA   VT   WA   WI   WV   WY <NA> 
  29    1    5    3   45   44   11   14    3   70   12    1    2  513 

---------------------------

Value counts for: self_employed 

  No  Yes <NA> 
1091  142   18 

---------------------------

Value counts for: family_history 

 No Yes 
762 489 

---------------------------

Value counts for: treatment 

Not Treated     Treated 
        619         632 

---------------------------

Value counts for: work_interfere 

    Never     Often    Rarely Sometimes      <NA> 
      212       140       173       464       262 

---------------------------

Value counts for: no_employees 

           1-5        100-500         26-100       500-1000           6-25 
           158            175            288             60            289 
More than 1000 
           281 

---------------------------

Value counts for: remote_work 

 No Yes 
880 371 

---------------------------

Value counts for: tech_company 

  No  Yes 
 226 1025 

---------------------------

Value counts for: benefits 

Don't know         No        Yes 
       407        371        473 

---------------------------

Value counts for: care_options 

      No Not sure      Yes 
     499      313      439 

---------------------------

Value counts for: wellness_program 

Don't know         No        Yes 
       187        837        227 

---------------------------

Value counts for: seek_help 

Don't know         No        Yes 
       363        641        247 

---------------------------

Value counts for: anonymity 

Don't know         No        Yes 
       815         64        372 

---------------------------

Value counts for: leave 

        Don't know Somewhat difficult      Somewhat easy     Very difficult 
               561                125                265                 97 
         Very easy 
               203 

---------------------------

Value counts for: mental_health_consequence 

Maybe    No   Yes 
  476   487   288 

---------------------------

Value counts for: phys_health_consequence 

Maybe    No   Yes 
  273   920    58 

---------------------------

Value counts for: coworkers 

          No Some of them          Yes 
         258          771          222 

---------------------------

Value counts for: supervisor 

          No Some of them          Yes 
         390          349          512 

---------------------------

Value counts for: mental_health_interview 

Maybe    No   Yes 
  207  1003    41 

---------------------------

Value counts for: phys_health_interview 

Maybe    No   Yes 
  555   496   200 

---------------------------

Value counts for: mental_vs_physical 

Don't know         No        Yes 
       574        338        339 

---------------------------

Value counts for: obs_consequence 

  No  Yes 
1070  181 

---------------------------

Value counts for: age_group 

10–15 16–20 21–25 26–30 31–35 36–40 41–45 46–50 51–55 56–60   60+ 
    0    22   195   362   339   185    92    30    12    10     4 

---------------------------

Country breakdown

# Count responses by country
country_counts <- mental_health %>%
  count(country, sort = TRUE)

# Calculate percentages
country_counts <- country_counts %>%
  mutate(
    pct = n / sum(n),
    pct_label = paste0(round(pct * 100, 1), "%")
  )

# Generate packed circle layout
packing <- circleProgressiveLayout(country_counts$n, sizetype = 'area')

# Combine layout with data
country_counts <- bind_cols(country_counts, packing)

# Generate circle vertices for plotting
circle_data <- circleLayoutVertices(packing, npoints = 50)

# Create a label column for text + tooltip
country_counts <- country_counts %>%
  mutate(
    label = ifelse(n > 3,
                   paste0(country, "\n", n, " Responses\n", pct_label),
                   "")
  )

# Create plot
p_bubble <- ggplot() +
  geom_polygon(data = circle_data,
               aes(x, y, group = id, fill = as.factor(id)),
               color = "white", alpha = 0.7, show.legend = FALSE) +
  geom_text(data = country_counts,
            aes(x = x, y = y, label = label),
            size = 3, color = "black", fontface = "bold", lineheight = 0.9) +
  coord_equal() +
  theme_void() +
  labs(title = "Survey Responses by Country (Packed Bubble Chart)")

# Make interactive with proper tooltip
ggplotly(p_bubble, tooltip = "label")

59.6% of the survey responses come from the USA with United Kingdom (14.7) and Canada (5.7%) being the next two largest. This introduces bias as how americans interact with their mental health in the workplace may be different to how other european countries work as well as eastern countries. ### Age Distribution

p1 <- ggplot(mental_health, aes(x = age)) +
  geom_histogram(binwidth = 5, fill = "steelblue", color = "black") +
  labs(title = "Age Distribution of Survey Respondents",
       x = "Age",
       y = "Count")
ggplotly(p1)

Most of the dataset is people whos age range from 20 to 40. This means the tech sector is mostly young workers in their careers.

Gender Distribution

p2 <- ggplot(mental_health, aes(x = gender, fill = gender)) +
  geom_bar() +
  labs(
    title = "Gender Distribution of Survey Respondents",
    x = "Gender",
    y = "Count",
    fill = "Gender"
  ) +
  theme_minimal()
ggplotly(p2)

There is significantly more men than women in the dataset which is to be expected as tech is a male dominated field.

Correlation Heatmap with treatment

# Select variables and convert to numeric
cor_data <- mental_health %>%
  select(
    work_interfere,
    age,
    treatment,
    remote_work,
    tech_company,
    care_options,
    wellness_program,
    seek_help,
    anonymity,
    leave,
    mental_health_consequence,
    phys_health_consequence,
    coworkers,
    supervisor
  ) %>%
  mutate(across(everything(), ~ as.numeric(as.factor(.))))  # convert all to numeric

# Compute correlation matrix
cor_matrix <- cor(cor_data, use = "complete.obs") %>% round(2)

# Interactive heatmap with heatmaply
heatmaply(
  cor_matrix,
  main = "Interactive Correlation Heatmap of Workplace Mental Health Factors",
  colors = colorRampPalette(c("#E46726", "white", "#6D9EC1"))(100),
  xlab = "", ylab = "",
  margins = c(60, 60, 40, 20),
  grid_color = "grey80",
  plot_method = "plotly"
)

1. Treatment by Age Group

p3 <- ggplot(mental_health, aes(x = age_group, fill = treatment)) +
  geom_bar(position = "fill") +
  labs(title = "Proportion of Treatment by Age Group",
       x = "Age Group",
       y = "Proportion") +
  scale_y_continuous(labels = scales::percent_format()) +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))
ggplotly(p3)

2. Treatment by Gender

p4 <- ggplot(mental_health, aes(x = gender, fill = treatment)) +
  geom_bar(position = "fill") +
  labs(title = "Proportion of Treatment by Gender",
       x = "Gender",
       y = "Proportion") +
  scale_y_continuous(labels = scales::percent_format())
ggplotly(p4)

3. Treatment by Country

top_countries <- mental_health %>%
  count(country, sort = TRUE) %>%
  top_n(10, n) %>%
  pull(country)

mental_health_top <- mental_health %>%
  filter(country %in% top_countries)

p5 <- ggplot(mental_health_top, aes(x = country, fill = treatment)) +
  geom_bar(position = "fill") +
  labs(title = "Proportion of Treatment by Top 10 Countries",
       x = "Country",
       y = "Proportion") +
  scale_y_continuous(labels = scales::percent_format()) +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))
ggplotly(p5)

4. Treatment vs Remote Work

p_remote_treatment <- mental_health %>%
  filter(!is.na(remote_work), !is.na(treatment)) %>%
  ggplot(aes(x = remote_work, fill = treatment)) +
  geom_bar(position = "fill") +
  scale_y_continuous(labels = scales::percent_format()) +
  labs(
    title = "Mental Health Treatment vs Remote Work",
    x = "Remote Work",
    y = "Proportion",
    fill = "Sought Treatment"
  )

ggplotly(p_remote_treatment)

5. Work Interference vs Mental Health Benefits

p_interference_treatment <- mental_health %>%
  filter(!is.na(work_interfere), !is.na(treatment)) %>%
  ggplot(aes(x = work_interfere, fill = treatment)) +
  geom_bar(position = "fill") +
  scale_y_continuous(labels = scales::percent_format()) +
  labs(
    title = "Treatment vs Work Interference",
    x = "Work Interference",
    y = "Proportion",
    fill = "Sought Treatment"
  )

ggplotly(p_remote_treatment)

6. Work Interference vs Mental Health Benefits

p_benefits_interfere <- mental_health %>%
  filter(!is.na(tech_company), !is.na(work_interfere)) %>%
  ggplot(aes(x = tech_company, fill = work_interfere)) +
  geom_bar(position = "fill") +
  scale_y_continuous(labels = scales::percent_format()) +
  labs(
    title = "Work Interference vs Mental Health Benefits",
    x = "Company Provides Mental Health Benefits",
    y = "Proportion",
    fill = "Interference With Work"
  )

ggplotly(p_benefits_interfere)

7. Leave Policy vs Seeking Treatment

p_leave_treatment <- mental_health %>%
  filter(!is.na(leave), !is.na(treatment)) %>%
  ggplot(aes(x = leave, fill = treatment)) +
  geom_bar(position = "fill") +
  scale_y_continuous(labels = scales::percent_format()) +
  labs(
    title = "Ease of Taking Leave vs Seeking Treatment",
    x = "Perceived Ease of Taking Medical Leave",
    y = "Proportion",
    fill = "Sought Treatment"
  ) +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))

ggplotly(p_leave_treatment)

8. Company Mental Health Resources vs Willingness to Talk to Supervisor

p_resources_vs_supervisor <- mental_health %>%
  filter(!is.na(care_options), !is.na(supervisor)) %>%
  ggplot(aes(x = care_options, fill = supervisor)) +
  geom_bar(position = "fill") +
  scale_y_continuous(labels = scales::percent_format()) +
  labs(
    title = "Awareness of Mental Health Care Options vs Willingness to Talk to Supervisor",
    x = "Aware of Mental Health Care Options",
    y = "Proportion",
    fill = "Would Talk to Supervisor"
  ) +
  theme(axis.text.x = element_text(angle = 45, hjust = 1))

ggplotly(p_resources_vs_supervisor)

Big Idea

Employers who provide strong workplace benefits—such as remote work flexibility, access to mental health resources, and supportive leave policies—create environments where employees are more likely to seek mental health treatment and report lower levels of work interference. This matters because untreated mental health issues not only harm individuals but also impact productivity, retention, and organizational culture. If companies recognize that supportive policies are not just perks but essential components of employee wellbeing, they can take actionable steps that improve both human and business outcomes.

Explanatory Visualizations